北京邮电大学学报

  • EI核心期刊

北京邮电大学学报

• 学术论文 • 上一篇    下一篇

基于业务管理的监控策略

    

  1.  北京邮电大学 电子工程学院 , 北京 100876
  • 基金资助:
     

Monitoring and Prediction Mechanism for Service Management 

    

  1. Electronic Engineering School, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Supported by:
     

摘要: 提出了一套灵活、有效的业务监视和预测方法 . 该方法利用时间相关性构造了业务模型 ,用离散傅里叶变换确定了合理的业务监测周期 ,并用贝叶斯公式对业务的异常行为进行预测 . 利用 一组网络实测数据证明了该方法的有效性 . 

关键词: 业务测量, 业务模型 , 性能预测

Abstract: A flexible and efficient method of service monitoring and prediction are presented. The time correlation is used to model the service behavior, the discrete Fourier transform is used to determine the reasonable monitoring period and the Bayesian theory is used to predict the abnormal behavior. In the end, a set of measurement data is used to show the validity of the method.

Key words: service measurement, service model, performance prediction

中图分类号: